Abstract
Ecological networks offer a comprehensive view of communities by capturing potential species interactions. While valuable for studying ecological change in the Anthropocene, many studies lack data across expansive temporal and spatial gradients. We addressed this gap by applying network approaches to paleolimnological records capturing strong land-use changes. We analyzed cladoceran assemblages, key aquatic organisms with identifiable subfossils, using two paleolimnological methods: (i) top-bottom comparisons of sediment records from 101 Canadian lakes with varying land-use intensity, and (ii) high-resolution core records from two impacted lakes in eastern Canada. We used correlation matrices of taxon relative abundances to calculate network metrics across land-use types and time periods. We found that lake communities currently experiencing high human impact changed through time, showing a decrease in connectance (proportion of realized to potential links) and an increase in modularity (measure of network subcommunities); these patterns were also observed in our full core analyses as well as in our randomized simulation exercise. Overall, this first pan-Canadian study of zooplankton paleo-networks provides new insights into how lake food webs have changed over a period of accelerated anthropogenic change.